4X on Fire: Prototyping

Prototyping
how a well-used model is improved upon

Simple:
Every example of a particular technology is exactly the same difficulty to build (ie an Apple IIe made in 1984 costs the same as an Apple IIe made in 1995).

Medium:
The first example of a new model is more expensive to build, the rest are the same. (a la Spaceward Ho!)

Complex:
    Each example of a particular model of a technical product costs less to build (at least manufacturing effort, possibly also materials) and increases the reliability of the product based on the data available from previous examples. There are two components which contribute to development of a model; data from real-world performance, and laboratory effort toward improvement.
    Data from performance is obviously not available until a space ship gets back (if the object is a ship) or reaches some research institution.   Data from battles or from encountering strange situations (almost running out of fuel, high radiation environment, high velocity relative to interstellar medium, etc.) is more valuable than any data which could be gotten in a groundside lab, and improves the development effort enormously. Sensor data from the performance of a ship will be worth partial credit (and is only available if the original ship is destroyed). Combat logs should be worth far more than normal operation logs. Such performance data might be transferred to computer ships each turn, and eventually carted back to your territory like ordinary cargo.
    Lab improvement is the refinement of a design based only on work in the laboratories of your settlements. It is essentially the equivalent of throwing money (and time) at the problem.
    The equation by which lab work and performance data are joined together to get total development level is:

D^aE^bor (kD)^a(gE)^b
    Where:
       D is the data from real-world performance
       E is the laboratory effort
       a, b, k, and g are numbers used to make the universe of one game different from that of another. Increasing a and k make real-world data more powerful, increasing g and b make laboratory work more valuable.

    The manufacturing effort used to create a product you've been making for a long time will be less than that for a new product. Your factories will be better attuned to making the product, as will any workers. The most efficient methods will be known. The equation for this improvement should be along the lines of

(dunno, let me think on it)

    The reliability of a well worked out product is obviously much higher than that of a new prototype. For example, a new prototype might only have a 50% chance of working correctly for a year, while a product you've been making for 1000 years might have a 99.999% chance. The equation for reliability should be along the lines of Reliability = (1 - x^Log(d)) where x is a constant between zero and one, and d is the total development level. Another option for this equation is the time-dilation form: Reliability = 1 / (1-d²/x²). Both of these give approximately the same form (I think) but the log(d) one is for if you have no theoretical maximum development quantity, and the d²/x² one is for where the "infinity" development level is x.

This work is licensed under a Creative Commons Attribution-NoDerivs 2.5 License.